Browsing by Author "Espadoto, Mateus"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    HyperNP: Interactive Visual Exploration of Multidimensional Projection Hyperparameters
    (The Eurographics Association and John Wiley & Sons Ltd., 2022) Appleby, Gabriel; Espadoto, Mateus; Chen, Rui; Goree, Samuel; Telea, Alexandru C.; Anderson, Erik W.; Chang, Remco; Borgo, Rita; Marai, G. Elisabeta; Schreck, Tobias
    Projection algorithms such as t-SNE or UMAP are useful for the visualization of high dimensional data, but depend on hyperparameters which must be tuned carefully. Unfortunately, iteratively recomputing projections to find the optimal hyperparameter values is computationally intensive and unintuitive due to the stochastic nature of such methods. In this paper we propose HyperNP, a scalable method that allows for real-time interactive hyperparameter exploration of projection methods by training neural network approximations. A HyperNP model can be trained on a fraction of the total data instances and hyperparameter configurations that one would like to investigate and can compute projections for new data and hyperparameters at interactive speeds. HyperNP models are compact in size and fast to compute, thus allowing them to be embedded in lightweight visualization systems. We evaluate the performance of HyperNP across three datasets in terms of performance and speed. The results suggest that HyperNP models are accurate, scalable, interactive, and appropriate for use in real-world settings.
  • Thumbnail Image
    Item
    Selecting and Sharing Multidimensional Projection Algorithms: A Practical View
    (The Eurographics Association, 2020) Espadoto, Mateus; Vernier, Eduardo Faccin; Telea, Alexandru C.; Gillmann, Christina and Krone, Michael and Reina, Guido and Wischgoll, Thomas
    Multidimensional Projection techniques are often used by data analysts for exploring multivariate datasets, but the task of selecting the best technique for the job is not trivial, as there are many candidates and the reasons for picking one over another are usually unclear. On the other hand, researchers developing new techniques can have a hard time comparing their new technique to existing ones and sharing their code in a way that makes it readily available for the public. In this paper, we try to address those issues systematically by analyzing recent surveys in the area, identifying the methods and tools used, and discussing challenges, limitations, and ideas for further work.